ANALYTICAL CHEMISTRY
1008 (3) Glendenin, L. E., Nucleonics, 2, No. 1, 12-33 (1948). (4) Glendenin, L. E., Science, 24, 623-6 (1950). (5) Hume, D. N., Glendenin, L. E,, and Ballou, N. E., Plutonium Project Report, CN-1312 (May 1945). (6) Jaffey, A. H., Kohman, T.p., and Crawford, J . A., Ibid., CC1602 (hlarch 1944). (7) Katcoff, S., Finkle, B., Sugarman, N., Glendenin, L. E., and
Winsberg, L., "Radiochemistry-the Fission Products," NNES Div. IV, 1'01. 9, Paper 7.35, ?Jew York McGraw-Hill Book Co., 1951. Livingston, h'. and Bethe, H. 266 ff. (1937).
Rev. Modern
Phus.,
9,
(9) Price, G. R., Feretti, R. S., and Schwarta, S., Plutonium Project Report, CC-2985 (June 1945). (lo) Peaborg, G. T.9 and Perlman, 1.3 Reo. Modem PhUs., 2% 585667 (1948). Symposium! Nuczeo71ics~ No* 3, 9-23 (1949). (11) waste (12) Yaffe, L., and Justus, K. M., J . Chenz. SOC.,1949, Supplementary Issue 2, 5341-51. (13) Zumwalt, L. R., Manhattan District Declassified Report, MDDC-1346 (1947). 47
24, 1950. Presented before the Division of Water, Sewage, a n d Sanitation Chemistry, Symposium o n Radioactive Waste Problems a n d Treatment, a t the 119th Meeting of the A > f E R I c A x CHEMICAL SOCIETT, Chirapo, 111.
Determination of Sulfur after Combustion in a Small Oxygen Bomb Rapid Titrimetric Method K. K. SIEGFRIEDT, J . S. WIBERLEY, AND R . W. MOORE Socony-Vacuum Laboratories, Brooklyn, N . Y .
C
HEMICAL procedures for the determination of sulfur in or-
ganic compounds generally consist of two separate phases. The first is the quantitative conversion of all sulfur into a single chemical form-Le., a sulfate or sulfide-and the second is the estimation of the amount of this sulfur product, Many combinations are possible. An analyst may alternately use any one of several method8 in one phase of the determination, while retsining a single technique for the other. Organic matter can be completely oxidized in a number of ways for the conversion of sulfur to inorganic sulfate. Wet oxidation with nitric acid according to Carius (11, 14, 28, 32), oxidation with perchloric acid (15),combustion in a tube in a stream of oxygen (12, 19, 28, 32, 40),,combustion in a lamp with air (4, 6, 43), fusion in a crucible with sodium carbonate or Eschka's mixture (3, 7 , 9 ) , fusion in a bomb with sodium peroxide (3, I S , 28, XI), and combustion in a bomb with oxygen under pressure ( 5 , 3 5 )are extensively used. Reduction methods, such a s catalytic hydrogenation (16, 17, 41, 42), have been less widely used. A few investigators have combined oxidation and reduction (22,25,37). The second phase of the sulfur determination also offers many possibilities. The gravimetric method based on the weighing of sulfur as barium sulfate is still the most commonly used in spite of the considerable expenditure of time, if the precipitate is allowed to stand before filtration (3-6,7 ) . Recently, a rapid gravimetric method was reported in which the sulfur trioxide is collected by silver gauze and weighed as silver sulfate (39). Volumetric as well as gravimetric methods employing benzidine for the precipit a n t have long been known (10). Simple acidimetric methods are sometimes applicable, but difficulties arise in the presence of other acid-forming elements, such az chlorine (6, 8, 28, 3 2 ) . Barium chromate methods have been described (18, 27), and the literature of the past 15 years contains many references to the titrimetric procedure which uses tetrahydroxyquinone as indicator (2, 9, 20,23,24, 26, 29, 31, 33-36,38,40, 43). The last method has been the subject of a great deal of discussion, particularly concerning the erne of detecting end points, relisbility in the hands of inexperienced operators, and the stoichiometry of the titration. When the product formed is sulfide, it may be determined iodometrically ( d l ) , oxidimetrically with hypochlorite (22),colorimetrically with methylene blue ( l 7 ) ,or by precipitation of a heavy metal (16). A method for general industrial analytical work should be ap-
A need for a small scale oxygen bomb and a more rapid method for the determination of sulfur led to the construction and testing of a 40-ml. bomb for analytical use. A convenient microprocedure for sulfur was devised which, by increasing the sample size, may also provide a rapid routine method for determining sulfur in petroleum products. Such a small bomb has been found suitable for the cornbustion of organic samples weighing from 2 to 200 mg. The bomb washings are titrated for sulfate with 0.02 .V barium chloride, using tetrahydroxyquinoneas indicator. It is not necessary to add at least 3 ml. of barium chloride, as some authors have supposed. A magnetic stirrer is advantageous in obtaining rapid end points and no difficulty has been experienced i n seeing the color change. Sulfur in the range of 0.01 to 100% may be determined without interference from most common elements. A single determination requires less than 1 hour of elapsed time with 20 to 30 minutes of operator time.
plicahle to the determination of high and low concentrations of sulfur in the presence of common metallic and nonmetallic elements. It should be rapid with respect to both operator time and total elapsed time and, if possible, should be suitable for microanalytical work. It seemed interesting to investigate the application of a small oxygen bomb, which a t that time was neither commercially available nor described in the literature. The method set forth in this article combines the oxidation in such a bomb with an improved titrimetric tetrahydroxyquinone procedure. APPARATUS
Oxygen Bomb. The bomb (Figure 1) has a capacity of approximately 40 ml., .and is so constructed that i t will prevent losses due to leakage of gases and permit quantitative collection of the combustion product with a minimum of wash liquid. The inner surface of the bomb may be made of stainless steel or any other material that will not be affected by the combustion process or the products of combustion. Materials used in the
10%
V O L U M E 23, NO. 7, J U L Y 1 9 5 1 bomb assembly, such as the head gasket and lead-wire insulation, should he resistant to heat and chemical action, and should not undergo any reaction which will affect the sulfur content of the liquid in the bomb. A Parr oxygen microhamh has been found satisfactory for this application. The platinum sample cup, 12 mm. in outside diameter, has a rim with an outside diameter of 20 mm. Over-all height 1s 5 mm.
I ~~
"
~
Arthur H. Thomas Co., Philadelphia, Pa. The titration stand should he capable of holding two burets, and have a white background illuminated with light such as that from a white fluorescent bulb. A glass color filter, Corning, with transmittance of approximately 37% a t 550 mp, may be obtained from Arthur H. Thomas Co., Philadelphia, Pit. (Catalog So. 9324-H). RRAGIIYTS
.. Ethyl shohol, 95%,
sulfur-free: Form& 30 or 3-A is acceptahle. Barium Chloride Standard Solution, 0.02 N. Dissolve 2.435 grams C.P. barium chloride dihydrate in water and dilute to 1 liter in a volumetric flask. Verify the normality of the solution by titrating 5-ml.portions of a standard sulfate solution prepared by dissolving 1.400 grams (weighed to the nearest milligram) of anhydrous sodium sulf%tein n,ater to give 1 liter of solut,ion. PROCEDURE
If the bomb is still wet from a previous rinsing, shhake the head to remove large droplets of water ahich might suhsequent,lpdrop into thesamplecupand prevent complete combustion. In the center of a piece of firing wire about 8 cm. in length, form a small coil of about seven loops I mm. in diameter. Connect the wire to the bomb terminals, arranging it in such a fashion that the coil t,onches the sample when the sample cup is properly placed in the bomb. Place several drops of watcr (approximately 0.5 ml.)
tain between 0.3 and 2.5 mg. of sulfur. If the sample is a solid or weighs less than 50 mg., add 1 drop of sulfur-free white oil to the sample in the cup. The total of sample plus white ail should not exceed 200 mg. Place the sample cup in position, assemble the bomb, and tighten the cover securely. Admit oxygen slowly (to avoid hlowing the sample from the cup) until a pressure of 500 to 550 pounds per square inoh is reached. It is essential that the homh does not leak under pressure. Leaks can he debected by filling the bomb with oxygen a t 500 pounds per square inch pressure and immersing it in water. If no bubbles appear, the bomb is correctly sealed. Connect the terminals to the electrical circuit and ignite the samde. Cool the bomb in runnine: water and rinse with distilled
mination. Wash the interior of the bomb thoroughly with a fine jet of d i e t,illed water and collect the rinsings in a 150-ml. beaker. Fifty milliliters of wash wat.er are usually sufficient. .Add approximately 3 ml. of saturated bromine water to the washings, place a cover glass on the beaker, and boil off the excess bromine. When the bromine has hailed off (several minutes' hoiling), add 2 drops of phenolphthalein and then 1N spdlum hydroxide dropwise until the solution is alklkaline. Permit the solution t o cool briefly and filter through qualitative filter paper into a 125-ml. Erlenmeyer flask. Rinse the heaker three times with 2 to 3 ml. of mater, pouring the rinsings through the filter paper. Wash the filter paper twice with 3 to 5 ml. of water. Evaporate the filtrate in the Erlenmeyer flask to 20 i 5 ml., cool to room temperature, then carefully neutralize with 0.05 N hydrochloric acid until the red phenolphthalein calor is just discharged. Add a measuring cup full of tetrahydroxyquinone indicator (approximately 100 mg.) and 20 ml. of95% ethyl alcohol. Place the Erlenmeyer flask eontaimng a stirring bar on the magnetic stirrer and arrange the orange filter near the flask, so that light from the titration stand can easily he seen passing through both. Titrate the contents of the flask with standard 0.02 N barium chloride while stirring vigorously. When 1 drop of barium chloride causes the color of the solution to become darker than the color of the filter and the darker color persists for 1 minute of vigorous stirring, the end point has been reached. \Then the tetrahydroxyquinone and the alcohol have been added, the solution has a yellow color. As barium chloride is added, the solution may momentrtrlly become red-orange, hut stirring rapidly changes the color beck t o yellow. At the end point, the solution is a dark orange, and the dark orange color will not fade even after several minutes' stirring. If the end paint has been passed, the solutio? will have a pronounced red hue. Make a blank test using the specified quantities of reagents and following the above procedure except for the combustion. If white oil is used, a combustion of the vhite oil should he made. Calculation. Calculate the sulfur content of the sample ea fallows:
% sulfur
=
1603 N ( A
-B]
II'
where
A B N W
= = = =
ml. of barium chloride used in titrating sample ml. of barium chloride useJ in titrating blank normalityof barium chloride weight of sample, mg. DISCUSSION
Oxygen Bomb. When this project was initiated, au oxygen bomb acceptable for microanalytical work was not available. A suitable bomb was designed and constructed hy the engineering and machine shop staff. The design of the customary Parr oxygen bomb 1102 was scaled down to one tenth, which gave the small bomb a capacity of 40 ml. Thus, when filled with oxygen to a pressure of 5M) pounds per square inch, it is suitable for samples of from 1 to 200 mg. in weight. Tests on the homh showed it t o be capable of oxidizing a variety of products. Pure organic compounds, petroleum products, deposits, ete., were burned without the formation of a carbon residue. Becamuse the sample size can be varied over a wide range, this apparatus SUffiC%S for samples containing a fen tenths of 1% of sulfur as well as for materials in which sulfur is a mzjor constituent. The bomb is sufficiently
1010
small for microanalytical nork, but is large enough to present no difficultyto the easy introduction of the sample and connection of the electrical firing wire. Shortly after completion of these studies, it was learned that the Parr Instrument Co. v a s independently perfecting a 5O-ml. oxygen microbomb for essentially the same purpose. This bomb has since been used by Agazzi, Parks, and Brooks in determination of sulfur and chlorine. In their procedure sulfur is determined gravimetrically ( I ) . One of these bombs was obtained on loan and evaluated with the same samples that had been previously analyzed. The results were found to agree within the limits of the method, so that the Parr bomb could be used interchangeably with the one the authors had constructed. 4 s a satisfactory oxygen microbomb is now available, plans of the bomb built a t the Socony-Vacuum Laboratories are not being published a t this time. Titration. The successful use of tetrahydroxyquinone in barium chloride titrations of sulfate depends upon the ease with which an operator can detect the end point. Pseudo end points cause the greatest concern. They arise from the slow precipitation of barium sulfate and the localized formation of the red barium salt of the indicator, which tend to appear toxard the end of the titration. Initially, after the indicator and the alcohol have been added, the solution is pale yellow. The addition of barium chloride roduces tiny patches of the red salt of the indicator, but tfese quickly disappear as the solution is agitated. -4sthe end point is approached, the yellow colo’r gradually gives way to an orange hue. At this stage, when most of the sulfate has been precipitated, further addition of barium chloride requires a longer time to react with the remaining sulfate. Because barium ions in solution will combine with the indicator to form a bright red salt, insufficient shaking will lead to the selection of a premature end point. The true end point is reached when the solution has a persistent dark orange color. Further additions of barium chloride will produce intense red shades. One may conclude, therefore, that rapid stirring is essential to the successful completion of a titration. Consequently, a magnetic stirrer is a great improvement over shaking or stirring by hand, and reproducible end points are quickly obtained without imposing any strain on the operator. An orange color filter provides a convenient means of repeat edly arriving a t the proper shade of color and thereby selecting identical end points. Kumerous titrations have shown that it is possible for different operators to reproduce end points with a precision of 0.02 ml. of 0.02 S barium chloride. Of considerable importance is the question of the stoichiometry of the titration. Ogg et al. (29) have reported that the use of a t least 3 ml. of 0.02 S barium chloride is necessary for a satisfactory titration. They state that, if less than 3 ml. of barium chloride is used in the standardization, the apparent normality will show rapid changes with corresponding small changes in the total volume of standard barium chloride solution used. Their curve has been reproduced as curve A in Figure 2. Actually, the normality does not change as curve A would lead one to believe. The shape of the curve is easily explained by taking into consideration an indicator blank. The authors’ work with zero concentrations of sulfate showed that 0.12 ml. of standard barium chloride solution was consistently consumed in each such titration. This amount was interpreted as the indicator error-Le., the quantity of barium chloride needed to produce a perceptible change in the color of the titrated solution. Several observers have pointed this out (20, SS-36, 40). Because 0.12 ml. is a significant volume of standard solution, it should be subtracted from the total volume of standard solution consumed in any titration. The effectof this indicator error on the apparent normality can be seen by referring to Figure 2. Curve B shows the change of the apparent normality of the barium chloride solution when standardized against 1-, 2-, 3-, 4-,5-, 6-, and 10-ml. portions of 0.0200 h’ sodium sulfate solution without consideration of the indicator blank. If only 1 ml. of barium solution is used in the titration, the indicator error accounts for about 10% of the
ANALYTICAL CHEMISTRY total volume of standard solution required. The resulting normality, therefore, must appear low by 10%. With larger totalvolumes of barium chloride solution, the percentage error will diniinish and the apparent normality will seem t o approach a constant value, as shown by curves A and B . This behavior disappears, as illustrated by curve C, if the buret readings are corrected for the indicator error. The normality then appears constant within 1% over the whole range from 1 to 10 ml., and the requirement may be dropped that more than 3 1111. of ptandard solution must be used in titrations,
4
a
0
3
I a
o
c
.
0
2
2
0
“ C U R V E A”
7
0.0210
a
m 0.0200
3
< z t
0.0190
0
b C U R V E 8”
W
a
2 L a 0.0170
0
2
4 6 8 IO MILLILITERS O F BARIUM CHLORIDE
Figure 2. Apparent Normality of Barium Chloride There is a close resemblance between curve -4obtained by Ogg et al. and curve B calculated from the aut,hors’data, Thus, it appears that failure to use a correct’ionfor the indicator blank leadr to exactly the same deviations as observed by Ogg et al., who seem to have missed this important point. They claim that standardizing the barium chloride with a standard sulfate solution “eliminates a correction factor.” They fail to observe, however, that it is the neglect of the indicator blank which produces the peculiar behavior of their apparent norm:dity. Furthermore, t,heir normality does not become constant xvhen more than 3 nil. of barium chloride solution are used for t,itration; their apparent normality increases 1%from 3 to 5 ml. and again 1% from 5 to 10 ml. of barium chloride used in titration. The authors’ assumption that the indicator error is constant, n-hich may not be strictly true (,% never I) leads , to results that are in error by inore t,han 1% throughout the range of 0 to 10 ml. of barium chloride solution. Because Ogg et al. do not provide for t’he determination of a blank, there is no safeguard against the effects of impurities in the reagents. Sulfate contaminations have been found, and it is believed that a blank determination should always be made. For the reasons given above, the frequently cited procedure of Ogg et al. is not recommended. Interferences have been treated by Sheen and Kahler (34). Barium contained in the sample precipitates an equivalent amount of sulfate, so that less barium chloride is needed in the titration. If the barium content is known, a correction may be applied by increasing the percentage of sulfur by 0.23 times the percentage of barium. Phosphorus interferes a t the p H specified for the titration, but moderate amounts of phosphorus, up to 60 p.p.m. in the final solution, can be tolerated in the titration, if the pH is reduced to 4.0. Lead compounds should not be burned in the platinum cup. Results on samples of both high and low sulfur contents are shown in Table I. Emph has been placed on samples Ion- in sulfur, because the primary objective was t,he development of a more rapid procedure than that of the ASTM (6). I t may be seen that the tetrahydro~;?-quiiioii[,titration provides adequate
~
V O L U M E 23, N O . 7, J U L Y 1 9 5 1
1011
Table 1. Comparison of Sulfur Results Obtained by Titrimetric-Oxygen Bomb Method with Theory or ASTM Values Sample
Comparison Method and Value
hlotor oil
D 894-48T
Cutting
D 129-44
0.18 0.21 0.19
n 01:
Xlotor oil
Heavy lube oil
Diesel oil
Cutting oil
D 894-48T
D 8o4-48T
n
EY-U
IR
0.49 0 51 0.49 0.50 0.62 0 6fi
0 86 0 8Y 0 88 0 88
0.87 0.89
1.16
1.11 1 12
1 22 1 48 1.4U
1
Fuel oil
1) 12SJ-44
0 0 .. 22 00 0.20 0.20
0.51 0.51 0.47 0 50 0 60 0.60 0 GO 0 60
1,16 I 14 1 , 13
D 128-44
TitrimetricOxygen Bomb Result
io
1 80 1.78 1 84 1 81
0.6.5 0 64
0.8.5 0.8;
1.10 1 11
1 47 1 . .5(1 1 ,;(I 1 49
I.i H 1.i 8 1.80 1 79
leum Products and Lubricants.” hlethotl D 90-47T, ASTM Committee D-2, 1949. Ibid.i Method 129-49. (6) Ibid., “Proposed Method of Test for Sulfur in Petroleuni Products by the Carbon Dioxide-Oxygen Lamp Method.” p. 1330. (7) Bssoc. Offic. Agr. Chemists, “Official and Tentative Methods of Analysis,” p. 126, 1945. (8) Brewster, E. L., and Rieman, Wm., 111, IND. Esc,. CHEM., AN.4L. ED.,14, 820 (1942). (9) Brunjes. H. L., and Manning, AI. J., Ibid., 12, 718 (1940). (IO) Callan, T. P., and Toennies,G.,Ibid., 13,460 (1941). (11) Carius, L., Ann., 116, l(1860). (12) Dennstedt, >I., Ber., 30,1590 (1897). (13) Elek, A., andHill, D. W., J. Am. Cheni. Soc., 55, 3479 (1933). (14) Emich, F., and Donau, J., Monatsh, 30, 764 (1909). (15) Evans, R. J., and St. John, J. L., IND.ENG.C m x . A N ~ LED., . 16, 630 (1944). (16) Field, E., and Oldach, C. S., Ibid., 18, 668 (1946). (17) Fogo, J. K., and Popowsky, hI., ANAL.CHEW,21, i34 (1949). (18) Foster, bI. D., IND. ENG.CHEM.,ASAL.ED.,8, 195 (1936). and Krekeler, H., -4ngew. Chem., 46, 103 (1933). (19) Grote, W., (20) Hallett, L. T., and Kuipers, J. W., IND. ENG.CHEM.,ANAL.ED., 12,360 (1940). (21) Heinemann, G., and Rahn, H. IT.,Ibid., 9,458 (1937). (22) Kirsten, TV., Mikrochemie, 35, 174 (1950). (23) Kochor. S.J..hD.EXG. CHEM. . AN A L . E D . . 9.381 (1937). (24) Lee, 9. k..Ti’allace, J. H., ,Jr,, Hand. IT. C’., and Hannay. S . B.,
.
Ihid.. 14. - 839 - - - (1942). 125) Luke, C . L., Ibid., 15, 602 (1943) (26) Mahoney, J. J., and hlichell, J. H., f b r t l . 14, 97 (1942) (27) Jlanov. G. C.. and Kirk, P. L., Ibid , 9 , 198 (1937). ( 2 8 ) Siederl, J . B., and Xiederl, Y,, ”Micromethods of Quantitative I
\----,
(1) Agassi, E. J., Parks, T. 11..:ind Brooks, E’. R.. . ~ - I L .C’HEJI., 23, (1951). (2) Alicino, J. F., Zbid., 20, 85 (1948). (3) =Im. Soc. Testing Materials. Philadelphia, “Book of .\.d.T..\I, Standards,” Part 5 , Method D 271-48, 1949.
Organic Analysis,” New York, ,John Wiley h Sons, 1942. (29) Ogg, C. L., Willits, C. O., and Cooper. F.J., ds.\r..CHEM.,20, 83 (1948). (30) Parr, 9. TT~,J . Ind. E u g . Chcm., 11, 230 (1919). (31) Peabody. W,A., and Fisher, R. S.,ISD.Esc. CHEX.,ASAL.ED., 10, 651 (1938). (32) Pregl, F., “Quantitative Organic Microanalysis.” Philadelphia, Blakiston’s Son & Co., 1930. 133) Schroeder. W. C.. IND. ENG.CHEM.. -\N.LL. ED..5. 403 (19331. (34) Sheen, R. T., andKahler. H. L.. Zbid., 8, 127 (1936). (36) Ibid.,10, 206 (1938). 1.36) Sheen, K.T.. Kahler. H. L.. and Cline, D. C., Ihid., 9, 69 (1937). ( 3 7 ) Sowa, E’. .J., -1rcadi. Y.(2.. and Xieuwland. .J. A , , I b i d . , 8, 49 (1936). ( 3 8 ) Steyermark. All,Bass. E.. and Littnian, B., -IN.%L. C H E Y . . 20, 5 8 i (1948). (39) Stragalid. G. L., and Safford, H. K., Ibid., 21, 625 (1949). (40) Sundberg, 0. E.. and Royer. G . L., I s n . Esti. (’HEM.. . i n . a ~ED., . 18, 719 (1946). (41) Ter JIeulen, H., and Heslinga, J.. “ X e u c Methoden der organischcheniischen Analyse.” Leipzig, Akademische Verlagsgesellschaft, 1927. (42) Wieseiiherger, E., Mikroclumie, 29, 73 (1941). E s c . (?HEX., ;ISAL. ED., (43) IYilson, C . T..and Kemper. W.A , . IND. 10,418 (1938).
(4) .\m, Soc. Testing J1ateri:tls. Philadelphia, “Standards on Petlo-
R E C L I V ED x e r n b e r 11, 1950.
Fuel oil
D 129.44
2 42 2 44 2.41 2.42
Cy.5tine (Bureau of Standards) Sulfur (U.S.P.)
Theory
26.i
Theory
Over 9 9 . 5
2.40 2.42 “39 1.40 26 G 100.0 99 2 49.6
accuricy. .4 single determination requires less than 1 hour of total elapsed time, 20 to 30 minutes of IT-hich are operator time. In routine use about sixteen determin:ttions per operator per day might he expected. LITER.iTITRE CITED
Determination of Sulfur and Chlorine in Organic Materials Reduced Scale Oxygen Bomb ELlGIO J. AGAZZI, THO\I IS 11. P4RKS1, *\D FRANCIS R. BROOKS, Shell Deuelopment Co., Emerycille, Calij.
T
H E sulfur or chlorine content of organic mateiials is generally determined in microanalysis by combustion methods such as the Sundberg and Roper (6) modification of the method of Grote and Krekeler ( 3 ) . These methods are widely applicable and generally satisfactory. However, in order to increase their sensitivity for materials t h a t have a low sulfur or chlorine content, the sample size must be increased proportionally. Proper quartztube combustion of large (25 to 50 mg.) samples is in many cases time-consuming and requires considerable care. Present address, Stanford Research Institute, Stanford, Calif.
I t was believed that utilization of a n oxygen bomb for the combustion of such samples might be advantageous, inasmuch as the combustion, which is carried out in a great excess of oxygen under pressure, requires very little operator time and attention. Sulfur and chlorine are determined using a n oxygen bomb by standard methods of the Bmerican Society for Testing Materials ( 1 , 2). However, the standard Parr bomb employed has a capacity of at least 300 ml., and washing the products of combustion from the bomb requires 300 to 400 ml. of water, which makes evaporation to a volume suitable for the determination of small amounts of